Use dateutil in order to support parsing timestamps that include milliseconds

[simonbaird]

See https://projects.engineering.redhat.com/browse/ERRATA-9192 .

I believe just adding a '%f' to the format string is enough to get it working. You could consider trying to make it work with either format to make the transition easier, maybe with a wrapper function.

Question: Is there any particular reason why datetime.datetime.strptime is used in some places and time.strptime in others? Could it be made consistent for all the timestamp parsing?

Actually we might still change Errata Tool back to use the old format in the short term, which would mean this isn't needed.

import time
import datetime

short_format = "2019-12-05T00:00:00"
long_format = "2019-12-05T00:00:00.000Z"

#d = short_format
d = long_format

print(time.strptime(str(d), '%Y-%m-%dT%H:%M:%S.%fZ'))
#=> time.struct_time(tm_year=2019, tm_mon=12, tm_mday=5, tm_hour=0, tm_min=0, tm_sec=0, tm_wday=3, tm_yday=339, tm_isdst=-1)

print(datetime.datetime.strptime(str(d), '%Y-%m-%dT%H:%M:%S.%fZ'))
#=> 2019-12-05 00:00:00

[simonbaird]

Jon has a WIP patch for Errata Tool to restore the old timestamp format, so I won't spend more time on this.

[jorris]

Or use dateutil and not have to worry about trivial format changes, especially when you aren't using the time component.
dateutil.parser.parse('2019-12-05T00:00:00.000Z').date().strftime('%Y-%b-%d')
'2019-Dec-05'

[ktdreyer]

Great idea jorris. I'm going to re-title this to use dateutil.